Representation of the impact of antibiotic administration on the bacterial community of the colon. After the onset of treatment, an increase in resistant bacteria (purple rods) can be seen. This increase is due to either a susceptible bacterium (green rods) becoming resistant or resistant bacteria, already present in low levels, increasing in number due to their ability to survive the selective pressure provided by the antibiotic. The acquired resistance is often due to horizontal gene transfer or mutation events (white arrow). As a consequence of treatment, a temporary decrease in diversity can also be seen. Some bacteria may be protected from antibiotic exposure in the mucin layer (yellow shading) or in grooves in between the villi formed by host epithelial cells that line the intestinal channel (not shown). The figure is not drawn to scale and the timescale is relative. 

Contexts in source publication

Context 1

... spread of resistant bacteria and resistance genes depends on different factors but the major pressure is antibiotic usage. Additional factors include the ability of resistant strains to colonize the gut, their relative fitness, mutation rates and efficiencies of horizontal transfer of resistance genes. Under the selective pressure of an antibiotic, a bacterium that acquires a resistance gene is often conferred with a benefit. When this selective pressure is no longer present, the resistant strain could have a lower fitness compared with its susceptible counterpart. However, this less competitive clone might compensate for this loss of fitness by acquiring compensatory mutations. The review by Andersson & Hughes (2010) discusses the fitness costs and mechanisms by which the bacterium can reduce these costs. A few studies have investigated the impact of antibiotics on long-term persistence of antibiotic resistance. The prevalence of erythromycin-resistant enterococci was investigated in subjects treated with clarithromycin (Sj ̈lund et al. , 2003). Using PFGE, it was shown that three of five subjects carried highly resistant enterococci clones 1 year post- administration and that these clones carried the ermB gene, conferring resistance to macrolides such as clarithromycin. In one patient, a specific resistant clone was detected 3 years after treatment in the absence of antibiotic pressure. In another study by Sj ̈lund et al. (2005) macrolide-resistant Staphylococcus epidermidis was detected up to 4 years after patients had taken clarithromycin. In our clindamycin study, we found significant increases in the levels of specific erm genes: ermF , ermG and ermB were detected in DNA extracted from the faecal samples by real- time PCR and these genes could still be detected 2 years after antibiotic administration (Fig. 3). Similarly, we observed long-term persistence of clindamycin-resistant Bacteroides clones following clindamycin treatment (Fig. 3) (Jernberg et al. , 2007). It was confirmed that most of the clones had acquired specific erm resistance genes. The initial cost of acquired resistance was high judging by slower growth in a culture of a resistant clone in competition with a susceptible clone (L ̈fmark et al. , 2008). However after 2 weeks, no growth disadvantage was detected and the subsequent resistant isolates that were collected retained their fitness up to 18 months after antibiotic treatment. Karami et al. (2008) also found no fitness burden in ampicillin-resistant Escherichia coli isolates compared with susceptible isolates from children up to 1 year old. These findings have important clinical implications, for example by providing extended opportunities for transmission of resistance genes to other species. Few additional studies have looked at the long-term selection and stabilization of resistance genes within bacterial populations. Recently, the pharyngeal carriage of streptococci and the proportion of macrolide-resistant isolates was studied in healthy volunteers after exposure to either of the two macrolides azithromycin or clarithromycin over 180 days (Malhotra-Kumar et al. , 2007). An increase in resistant strains was seen for both groups compared with a placebo group immediately after treatment. The proportion of resistant isolates was higher after azithromycin treatment than after clarithromycin use, while clarithromycin selected for ermB genes, which could not be seen with the azithromycin treatment. The results implicate that macrolide use is the major driving force of macrolide resistance. There is a high level of transfer of resistance genes within the intestine, as shown by several studies, but the picture is far from complete (Lester et al. , 2006; Salyers et al. , 2004; Scott, 2002; Shoemaker et al. , 2001). The intestine is apparently an ideal location for efficient transmission of resistance genes. This moist, warm environment with nutrients in abundance is comprised of high numbers of bacterial cells that are potential targets for resistance development and that also constitutes a large gene pool for resistance determinants (Fig. 1). After initial selection of resistance genes in the commensal microbiota, they may then potentially be transferred to pathogens. This is exemplified by a study that demonstrated transfer of a plasmid carrying a b -lactamase gene from a resistant E. coli strain to an initially susceptible strain in a child treated with ampicillin (Karami et al. , 2007). The authors con- cluded that antibiotics may not only select for resistant bacteria but also consequently increase the number of transfer events from the increased pool of resistant cells. Sommer et al. (2009) characterized the functional resistance reservoir in two unrelated healthy subjects who had not been exposed to antibiotics for at least 1 year. The microbiota was analysed using a metagenomic approach in addition to culturing. Sequencing of clones conferring resistance to 13 different antibiotics revealed 95 unique inserts that were evolutionarily distant from known resistance genes. This diverse gene pool of resistance genes in the commensal microbiota of healthy individuals could also potentially lead to the emergence of new resistant pathogenic strains. Some bacteria are only transient inhabitants of the intestine and are resistant to colonization, such as many that originate from ingested foods. However, they can still play a key role in the introduction of resistance determinants that have the potential to be transferred to the commensal microbiota in the intestine during passage (Andremont, 2003; Salyers et al. , 2004). Another factor that might be contributing to the emerging resistance problem is the use of antibiotics or analogous compounds in agriculture. The use of these compounds in agricultural settings may lead to a more constant selective pressure for resistance to develop and could potentially contribute to a larger global resistance reservoir with potential introduction, for example via opportunistic pathogens, into the clinical environment (Aubry-Damon et al. , 2004; Heuer & Smalla, 2007). Increasing antimicrobial resistance is a growing threat to human health and is mainly a consequence of excessive use of antimicrobial agents in clinical medicine. In addition to focusing on clinically relevant pathogens when monitoring levels and risks for emergence of antimicrobial resistance, it is important to also consider the role of the enormously diverse human commensal microbiota. It is generally acknowledged that the use of antibiotics causes selection for and enrichment of antimicrobial resistance, but it has also been believed until recently that the commensal microbiota is normalized a few weeks following withdrawal of the treatment. As discussed above, increasing evidence suggests that this is not the case and that specific members of the microbiota may be positively or negatively affected for extended periods of time. With the onset of improved sequencing methods and other molecular approaches, increasing information is becoming available about how the biodiversity (richness and evenness) of the human microbiome is affected at different phylogenetic levels, from the community level to species, strains and individual clones. At the same time, it is possible to monitor the acquisition and persistence of resistance genes in the community. Together this information should help to provide knowledge of the natural dynamics of the normal microbiota and help us to understand the long-term consequences of antimicrobial treatment. This information is of great importance for the implementation of rational administration guidelines for antibiotic ...

Context 2

... therapy can affect not only the target pathogen but also commensal inhabitants of the human host. The extent of the impact on non-target microbial populations depends on the particular antibiotic used, its mode of action and the degree of resistance in the community. Sometimes an imbalance in the commensal gut microbiota due to antibiotic administration can result in intestinal problems, such as antibiotic-associated diarrhoea (AAD) (McFarland, 1998). An additional concern is the increase in antibiotic resistance and the potential spread of resistance genes to pathogenic bacteria. Recently, it has been shown that even short-term antibiotic administration can lead to stabilization of resistant bacterial populations in the human intestine that persist for years (Jakobsson et al. , 2010; Jernberg et al. , 2007; L ̈fmark et al. , 2006). Although the consequences of long-term persistence of antibiotic resistance in the human gut are currently unknown, there are high risks that this could lead to increased prevalence of antibiotic resistance, reduce the possibility of successful future antibiotic treatments and subsequently lead to higher treatment costs. The short-term consequences of antibiotic administration have previously been reviewed and have primarily dealt with culture-based analyses. This mini-review will focus on the long-term consequences of antibiotics on the composition, ecology and resistance of the human gut microbiota and will highlight some recent studies based on molecular methods. Practically all surfaces of the human body exposed to the environment are normally inhabited by micro-organisms. The intestine constitutes an especially rich and diverse microbial habitat. Approximately 800–1000 different bacterial species and . 7000 different strains inhabit the gastrointestinal tract (B ̈ckhed et al. , 2005). These bacteria act together in many physiological processes and also interact with human cells, including those of the immune system. The diversity of the gut microbiota is relatively simple in infants but becomes more complex with increasing age, reaching a high degree of complexity in adults (Fanaro et al. , 2003). Lifestyle factors (Dicksved et al. , 2007) and diet (Flint et al. , 2007; Ley et al. , 2005, 2006) can also affect the diversity and composition of the gut microbiota. Interestingly, the relative proportions of the two most dominant bacterial phyla i.e. Bacteroidetes and Firmicutes , were found to be correlated with obesity in mice and humans, respectively (Ley et al. , 2005, 2006), but a study by Duncan et al. (2008) did not see this same correlation in obese and lean humans. Molecular analyses have also revealed that the composition of the human intestinal microbiota is host-specific (Dicksved et al. , 2007; Jernberg et al. , 2007) and relatively stable over time (Jernberg et al. , 2007; Zoetendal et al. , 1998). Recent metagenome sequencing data of 124 individuals suggest the existence of a common core human gut microbiome (Qin et al. , 2010), but this core may exist more at the level of shared functional genes rather than shared taxa (Turnbaugh et al. , 2007, 2009). The ecological balance between the human host and associated micro-organisms described above can be disturbed by several factors, most dramatically by administration of antimicrobial agents. This perturbation mainly manifests as decreased colonization resistance of members of the commensal microbiota, which leads to varying states of disease as well as emergence of antibiotic-resistant strains (Fig. 1) (de la Cocheti ` re et al. , 2005; Sj ̈lund et al. , 2003). Short- term changes in the quantity and composition of the bacteria comprising the normal human flora as a response to antibiotic exposure have been extensively recorded (for example Sullivan et al. , 2001). However, only a few recent studies have investigated the long-term impacts of antibiotic administration, including development of resistance (Jakobsson et al. , 2007; Jernberg et al. , 2007; Lindgren et al. , 2009; L ̈fmark et al. , 2006; Nyberg et al. , 2007; Sj ̈lund et al. , 2003). Most studies to date that have addressed the impact of antibiotics on the intestinal microbiota have been performed using culture-based techniques. Disadvantages of culturing are that despite the use of specific selective media and anaerobic incubation conditions, there remains a substantial part of the microbiota, approximately 80 % (Eckburg et al. , 2005), that has not yet been cultured. On the other hand, several species representing the main groups of clinically important opportunistic bacteria can be routinely recovered on culture media, including members of the genera Bacteroides , Streptococcus , Enterococcus and Staphylococcus and the family Enterobacteriaceae . Cul- turing is still valuable because the pure cultures that are generated can be further analysed with respect to their physiology and antimicrobial susceptibility (Table 1). Furthermore, isolates can be subtyped to strain or clonal levels enabling specific strains or clones to be monitored over time or according to treatment (Jernberg et al. , 2007; Sj ̈lund et al. , 2003). The limitations of culture-based techniques can be largely circumvented by using molecular approaches (Brug ` re et al. , 2009). For example, community fingerprinting approaches based on 16S rRNA gene amplification and characterization, such as terminal-restriction fragment length polymorphism (T-RFLP), denaturing/temperature gradient gel electrophoresis (DGGE/TGGE) and 16S rRNA gene sequencing are useful for tracking temporal changes or perturbations in response to antibiotics (Fig. 2). More recently, second generation sequencing approaches (Hamady & Knight, 2009), such as 454 pyrosequencing using specific bar codes to identify samples (Andersson et al. , 2008), have provided more in depth information about the impact of antibiotics on specific phylogenetic groups of the gut microbiota (Dethlefsen et al. , 2008). Different antimicrobial agents can influence the normal gut microbiota in different ways. The extent of the antibiotic- induced alterations in the microbiota depends on several factors: i) the spectrum of the agent, ii) dosage and duration of treatment, iii) route of administration and iv) the pharmacokinetic and pharmacodynamic properties of the agent. For example, secretion of an antibiotic by intestinal mucosa, bile or salivary glands may subsequently interfere with the normal microbiota at different sites (Sullivan et al. , 2001). Other side effects of some antibiotics on the human host include disturbance of the metabolism and absorption of vitamins (Levy, 2000), alteration of susceptibility to infections (Levy, 2000) and overgrowth of yeast (Sullivan et al. , 2001) and/or Clostridium difficile (Edlund & Nord, 1993; Sullivan et al. , 2001). This review focuses on the impact of antibiotics from a bacterial perspective. Since there is considerable subject-to-subject variability in the composition of the gut microbiota among humans (Turnbaugh et al. , 2009) the investigation of the impacts of antibiotics is currently best assessed on an individual basis. For example, in three healthy individuals treated with ciprofloxacin, certain abundant taxa showed inter-individual variation in response to the antibiotic. The overall effect of the antibiotic on many taxa was also stronger in two of the three individuals (Dethlefsen et al. , 2008). Grouping of microbial community data from several individuals can result in loss of statistical significance or false-negative results. Measuring individual divergences from baseline levels after treatment can overcome the problem of subject-to-subject variability (Engelbrektson et al. , 2006). However, variations in the microbiota due to other external factors, such as diet or stress, can complicate the task of untangling the specific impacts of antibiotic treatments. Therefore, more studies investigating the extent of long-term natural fluctuations in gut microbial communities are needed to establish knowledge about the extent of baseline fluctuations over time. Several antibiotics are specifically active against anaerobic bacteria that dominate in the human intestinal microbiota. They play an important role in maintaining a healthy gut, such as producing extensive amounts of volatile fatty acids. Therefore, treatment with antibiotics that select against important groups of anaerobic bacteria can have substantial consequences for the resultant functional stability ...